How joncarre/network-analyzer works
Project type and purpose
• Network Analyzer is a full-stack, microservice-style application for capturing, processing, storing and AI-enhanced analysis of network packet data.
• The backend (Python/FastAPI) orchestrates four core services (capture, processing, database, AI) and exposes REST endpoints. The frontend (React + TailwindCSS) provides an interactive web UI including tables, charts and a chat interface.
File structure
• Top level:
• backend/ – FastAPI app
• frontend/ – React app
• backend/ subfolders:
• api/ – route handlers (ai_api.py, capture_api.py, database_api.py, processing_api.py)
• / – TShark integration
• processing/ – PCAP → data structures
• database/ – ORM models, SQLite schemas
• ai/ – Claude AI prompts & metrics
• tests/, benchmarks/
• run.py, main.py, config (env, .env.example)
• frontend/ structure:
• src/components, src/pages – React components for CapturePage, AnalysisPage, ChatInterface, PacketTable, AnomalyTable, SessionAnalytics
• src/services/api.js – HTTP client
• Tailwind + Vite config
Key README insights
• Microservices-style architecture: Capture, Processing, Database, REST API, IA Integration
• Tech stack: FastAPI, SQLite, TShark, Pandas, React, TailwindCSS, WebSocket/REST, Anthropic Claude
• User flow:
1) Via TShark → PCAP
2) Auto-processing → validate + parse → SQLite DB
3) UI visualizes stats + tables
4) Chat interface to AI for natural-language analysis